The dark horse of quantum computing
In the race to build a useful quantum computer, most of the attention goes to familiar names and familiar approaches: superconducting circuits, trapped ions, ever-larger counts of physical qubits. But a quieter line of research — one that a leading scientist calls "a dark horse in the race to quantum error correction" — has just cleared a milestone that its supporters have been chasing for years.
A collaboration spanning the University of Chicago's Pritzker School of Molecular Engineering, Harvard University, Stony Brook University's C.N. Yang Institute for Theoretical Physics, and the quantum computing company Quantinuum has demonstrated a complete, universal set of quantum operations built from non-Abelian anyons — strange quasiparticles whose strange behavior may one day make quantum computers far more resistant to errors. The work was published in the journal Nature, and the results were reported widely in the final week of September 2026.
The headline claim is striking: for the first time, researchers have shown on real quantum hardware that this topological approach can, in principle, perform any quantum computation — a property called universality. It is a proof of concept at small scale, not a working machine. But it moves topological quantum computing from a beautiful theory with partial demonstrations to a framework with every logical primitive in place.
What anyons are — and why they matter
To understand the result, it helps to start with the central headache of quantum computing: errors. Quantum bits, or qubits, are exquisitely sensitive to disturbance. A stray vibration, a tiny temperature fluctuation, a whisper of electromagnetic noise can corrupt a calculation. The standard answer is error correction — spreading each logical unit of information across many physical qubits, so that no single failure destroys it. But that redundancy is enormously expensive, and even the best schemes leave gaps: certain essential operations cannot be performed directly on error-corrected qubits and must be supplied by costly additional procedures.
Non-Abelian anyons offer a fundamentally different bargain. They are not ordinary particles at all, but collective excitations — patterns that emerge when many ordinary qubits are entangled together in just the right way. Information stored in anyons is naturally spread across the whole entangled system rather than concentrated in one place, giving it an inherent protection against the small local disturbances that plague conventional qubits.
The "non-Abelian" part refers to what happens when you move anyons around each other — an operation called braiding. With ordinary particles, swapping two of them twice returns everything to its original state. With non-Abelian anyons, the order of the swaps matters: braiding them in different sequences produces different outcomes. That path-dependence is exactly what you need to perform computations — the braiding itself becomes the logic gate.
It is an elegant idea, and it has been pursued for decades. The catch has always been universality: showing that braiding and related operations can produce not just some quantum gates, but every gate needed for arbitrary computation.
Braiding plus fusion: the key insight
The new work's central insight is that braiding alone is not enough — at least not for the simplest, most practical kinds of anyons. That lesson came from the team's own earlier experiment. In 2024, some of the same researchers used a Quantinuum trapped-ion computer to create anyons associated with a symmetry group called D4 — the rotations and reflections that leave a square unchanged. It was the first demonstration of this kind of non-Abelian order on quantum hardware, proving the exotic particles could be created and manipulated. But as team member Ruben Verresen, an assistant professor of molecular engineering at UChicago, later put it: "In that work, we didn't demonstrate that those emergent forces were enough to do quantum computation. That particular universe we created was not powerful enough."
For the new study, the team switched to a different symmetry, known as S3 — the rotations and mirror flips that leave an equilateral triangle unchanged. S3 is the smallest non-Abelian group, making it the simplest system that could plausibly work. The researchers created the corresponding anyons on Quantinuum's H2 trapped-ion processor by entangling 54 qubits into a carefully prepared quantum state.
The S3 system had the properties needed for universal computation — but only when braiding was combined with a second operation called fusion. In fusion, two anyons are brought together and the outcome is measured, yielding information about their combined state. The underlying concept dates back to theoretical work by Carlos Mochon, then a student of physicist John Preskill at Caltech, in the early 2000s. Turning that mathematics into an experiment that could run on real hardware took more than two decades of additional theoretical and experimental work.
Rather than ordinary two-level qubits, the team encoded their information in "topological qutrits" — units with three possible levels instead of two, stored in the shared, nonlocal state of separated anyons. By combining braiding and fusion operations, they demonstrated the three essential tools: one entangling gate produced through braiding, and two different kinds of measurement produced through fusion. Together, those operations can in principle generate any quantum operation — including ones that braiding alone could never achieve.
"We demonstrated a universal gate set — meaning that if you store information in these particles, and you move them around, you can do any quantum computation you might want to do." — Ruben Verresen, assistant professor of molecular engineering, UChicago Pritzker School of Molecular Engineering
Why magic states are such a big deal
One of the most consequential parts of the demonstration was also its most technical: the team showed that their anyons could directly prepare a "magic state." To appreciate why that matters, consider one of the dirty secrets of quantum error correction. Even with excellent error-corrected qubits, certain operations needed for universal computation cannot be performed directly. Quantum computers get around this by preparing special resource states — magic states — and consuming them to unlock the missing operations. But producing magic states of sufficient quality requires a purification process called distillation, which devours a huge fraction of a quantum computer's available qubits.
The anyon approach sidesteps that bottleneck. Because the topological encoding natively supports the full gate set, the team could prepare a magic state directly through their braiding-and-fusion operations, without the resource-intensive distillation step. As Henrik Dreyer, managing director and scientific lead at Quantinuum's Munich office, noted, this suggests that fault-tolerant computation might one day be possible with far fewer physical qubits than conventional approaches demand — a prospect that could dramatically change the economics of building large-scale quantum machines.
"Non-Abelian codes are a dark horse in the race to quantum error correction." — Henrik Dreyer, managing director and scientific lead, Quantinuum Munich
What this doesn't mean — yet
For all its elegance, the result comes with clear limits, and the researchers do not hide them. The demonstration ran at small scale — 54 qubits encoding a handful of logical qutrits — and none of the operations were performed fault-tolerantly, meaning the computation was not yet protected against errors in the way a real machine would require. The current experiments also use relatively deep circuits; the team notes that the operations could in principle be implemented with more efficient, constant-depth adaptive circuits, but that remains future work.
There is also a broader strategic question. The anyons in this experiment were not found in a material — they were synthesized from ordinary qubits on a conventional trapped-ion processor. That is both the achievement and the irony: the team proved that a commercial quantum computer can emulate the physics needed for topological computation, without waiting for physicists to discover or engineer a native material hosting non-Abelian anyons. Whether the long-term future belongs to emulated anyons on conventional hardware or to purpose-built topological materials is still an open question.
Even so, the conceptual advance is real. A framework that was theoretically promising but experimentally incomplete now has its full logical toolkit demonstrated on hardware. The dark horse has shown it can run the whole course — the next challenge is running it at a scale, and with a level of error protection, that matters for real problems.
Sources
